US20220188734A1 - Processing device, processing method, and storage medium - Google Patents
Processing device, processing method, and storage medium Download PDFInfo
- Publication number
- US20220188734A1 US20220188734A1 US17/408,696 US202117408696A US2022188734A1 US 20220188734 A1 US20220188734 A1 US 20220188734A1 US 202117408696 A US202117408696 A US 202117408696A US 2022188734 A1 US2022188734 A1 US 2022188734A1
- Authority
- US
- United States
- Prior art keywords
- data
- plan
- sets
- performance
- new plan
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012545 processing Methods 0.000 title claims abstract description 137
- 238000003672 processing method Methods 0.000 title claims description 12
- 238000003860 storage Methods 0.000 title claims description 5
- 238000004519 manufacturing process Methods 0.000 claims abstract description 135
- 238000011156 evaluation Methods 0.000 claims abstract description 77
- 239000000284 extract Substances 0.000 claims abstract description 12
- 238000013527 convolutional neural network Methods 0.000 claims description 5
- 230000002950 deficient Effects 0.000 claims description 2
- 238000012423 maintenance Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 description 10
- 230000007423 decrease Effects 0.000 description 9
- 238000012986 modification Methods 0.000 description 6
- 230000004048 modification Effects 0.000 description 6
- 230000008901 benefit Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 230000002349 favourable effect Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000000034 method Methods 0.000 description 4
- 230000006978 adaptation Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000015654 memory Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000007639 printing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/906—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06313—Resource planning in a project environment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
- G06Q10/06375—Prediction of business process outcome or impact based on a proposed change
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Definitions
- Embodiments described herein relate generally to a processing device, a processing method, and a storage medium.
- a production plan is pre-generated when producing a product. It is desirable to develop technology that can provide a user with data that is useful for generating the production plan.
- FIG. 1 is a schematic view illustrating a processing system according to an embodiment
- FIGS. 2A to 2H are schematic views for describing processing according to the processing system according to the embodiment.
- FIGS. 3A and 3B are schematic views for describing processing according to the processing system according to the embodiment.
- FIGS. 4A to 4D are schematic views for describing processing according to the processing system according to the embodiment.
- FIGS. 5A and 5B are schematic views illustrating processing results according to the processing system according to the embodiment.
- FIG. 6 is a schematic view illustrating processing results according to the processing system according to the embodiment.
- FIGS. 7A and 7B are schematic views illustrating processing results according to the processing system according to the embodiment.
- FIG. 8 is a flowchart illustrating a processing method according to the processing system according to the embodiment.
- FIG. 9 is a schematic view for describing advantages of embodiments.
- FIG. 10 is a flowchart illustrating a processing method according to the processing system according to the modification of the embodiment.
- FIG. 11 is a schematic view illustrating a hardware configuration.
- a processing device refers to a plurality of data sets.
- Each of the data sets includes previous plan data and performance data.
- the previous plan data are of a time series of a relationship between time and a target production volume of a previous plan.
- the performance data are of performance with respect to the previous plan.
- the device calculates a plurality of first evaluation values for a plurality of sets of the performance data by using distances between new plan data and each of a plurality of sets of the previous plan data.
- the new plan data are of a time series of a relationship between time and a target production volume of a new plan.
- the first evaluation values are of an evaluation as a prediction of performance with respect to the new plan data.
- the device inputs a new plan image to a first model and calculates a plurality of second evaluation values for the sets of performance data by using a classification result of the new plan image output from the first model.
- the new plan image is of the relationship between time and the target production volume of the new plan.
- the second evaluation values are of an evaluation as a prediction of the performance with respect to the new plan data.
- the device extracts at least one of the sets of performance data by using the first evaluation values and the second evaluation values.
- FIG. 1 is a schematic view illustrating a processing system according to an embodiment.
- FIGS. 2A to 4D are schematic views for describing processing according to the processing system according to the embodiment.
- the processing system 10 provides a user with data that can be utilized in a prediction of performance relating to a new production plan.
- the processing system 10 includes a processing device 1 , a memory device 2 , an input device 3 , and an output device 4 .
- the memory device 2 stores a data set that includes previous plan data and performance data for the previous plan data.
- the previous plan data is time-series data of the relationship between time and the target production volume at each time of a previous plan for some product.
- the performance data is time-series data of the relationship between time and the actual volume produced (the actual production volume) at each time for the product.
- the previous plan data is of the relationship between dates and the target production volume of the product at each date.
- the performance data is of the relationship between dates and the actual production volume of the product at each date.
- the memory device 2 stores multiple data sets.
- the products that are the objects of the plans and performance may be the same or different from each other between the data sets.
- FIGS. 2A to 2D are examples of previous plan data.
- the previous plan data A 1 to D 1 of FIGS. 2A to 2D respectively show the plans of the production of products A to D.
- FIGS. 2E to 2H are examples of performance data.
- the performance data A 2 to D 2 of FIGS. 2E to 2H respectively show the performance with respect to the previous plan data A 1 to D 1 of FIGS. 2A to 2D .
- the horizontal axis is time; and the vertical axis is the production volume.
- a data set that includes the previous plan data A 1 of FIG. 2A and the performance data A 2 of FIG. 2E is stored in the memory device 2 .
- a data set of the previous plan data B 1 of FIG. 2B and the performance data B 2 of FIG. 2F , a data set of the previous plan data C 1 of FIG. 2C and the performance data C 2 of FIG. 2G , and a data set of the previous plan data D 1 of FIG. 2D and the performance data D 2 of FIG. 2H are stored in the memory device 2 .
- the processing device 1 acquires new plan data.
- the new plan data is time-series data of the relationship between time and the target production volume at each time of a new plan for any product.
- the new plan data is of the relationship between future dates and the target production volume of the product at each date.
- the product that is the object of the new plan may be the same as or different from the products that are the objects of the previous plans and performance.
- the user uses the input device 3 to input data to the processing device 1 .
- the data that is transmitted from the processing device 1 is output by the output device 4 toward the user.
- the user uses the input device 3 to input the new plan data to the processing device 1 .
- the new plan data is stored in the memory device 2 ; and the processing device 1 may acquire the new plan data from the memory device 2 .
- FIG. 3A is an example of new plan data.
- the new plan data N 1 of FIG. 3A shows the plan of the production of the product A.
- the horizontal axis is time; and the vertical axis is the production volume.
- the processing device 1 extracts performance data that can be utilized as a prediction of the performance with respect to the new plan from the multiple data sets.
- the processing device 1 extracts by executing the following first processing and second processing.
- the processing device 1 calculates the distances between the new plan data and the multiple sets of previous plan data.
- the processing device 1 calculates the multiple first evaluation values of an evaluation as a prediction of the performance with respect to the new plan data for the multiple sets of performance data by using the calculated multiple distances.
- the processing device 1 calculates the distances between the new plan data N 1 of FIG. 3A and the multiple sets of previous plan data A 1 to D 1 of FIGS. 2A to 2D .
- the distance decreases as the similarity of the form of the change of the target production volume with respect to time increases.
- the performance data for the previous plan data for which the distance is obtained becomes more suitable to predict the performance with respect to the new plan data as the distance decreases.
- the processing device 1 generates a ranking by arranging the multiple sets of performance data in order from the shortest distance for the multiple distances of the multiple sets of previous plan data.
- the ranking of the performance data corresponding to the previous plan data is higher as the distance for the previous plan data decreases.
- the processing device 1 calculates the first evaluation values of the performance data according to the order of the ranking.
- a Euclidean distance, a Manhattan distance, a Chebyshev distance, a distance calculated by dynamic time warping (DTW), etc., can be used as the distance.
- the processing device 1 uses the new plan image and a first model.
- the new plan image is image data of the relationship between time and the target production volume of the new plan.
- FIG. 3B is an example of a new plan image.
- the new plan image n 1 of FIG. 3B is image data corresponding to the new plan data N 1 of FIG. 3A .
- the image of a chart in which the horizontal axis is time and the vertical axis is the production volume is used as the new plan image n 1 .
- the content shown in the image is arbitrary as long as the relationship between time and the target production volume is shown.
- the image may be a scatter plot showing the relationship between time and the production volume.
- the user uses the input device 3 to input the new plan image to the processing device 1 .
- the processing device 1 may generate the new plan image based on the new plan data.
- the first model outputs a classification result according to the input of the plan image.
- the first model includes, for example, a neural network, a model trained by a random forest, a model trained by a decision tree, or a model trained by a support vector machine. It is favorable for the first model to include a convolutional neural network (CNN). Each of the multiple sets of performance data are classified as some class.
- the classification result includes the probability of the plan image being classified as each of the multiple classes.
- the first model is pretrained using a previous plan image.
- the previous plan image is image data of the relationship between time and the target production volume of the previous plan for some product.
- FIGS. 4A to 4D are examples of previous plan images.
- the previous plan images al to dl of FIGS. 4A to 4D correspond respectively to the previous plan data A 1 to D 1 of FIGS. 2A to 2D and show plans of the production of the products A to D.
- the image of a chart in which the horizontal axis is time and the vertical axis is the production volume is used as the previous plan image.
- the performance data A 2 to D 2 of FIGS. 2E to 2H respectively show the performance with respect to the previous plans of the previous plan images al to dl.
- the first model is trained by using the previous plan image as input data and by using a value of the class of the performance data as a label.
- the first model is trained by using the previous plan image al of FIG. 4A as input data and by using a value of a class that includes the performance data A 2 of FIG. 2E as a label. Thereby, the first model is trained so that the image is classified as some class when the plan image is input.
- the class is set according to the form of the change of the target production volume with respect to time. For example, the same class is set for previous plans that have similar forms of the change of the target production volume with respect to time regardless of the target production volume and the production period.
- the processing device 1 inputs the new plan image to the first model and acquires a classification result output from the first model.
- the processing device 1 uses the classification result to calculate a second evaluation value for each set of performance data.
- the second evaluation value is of an evaluation as a prediction of the performance with respect to the new plan data of the new plan image for the performance data.
- the processing device 1 generates a ranking by arranging the classes in order from highest probability. As the probability increases, the likelihood of the performance data classified as each class being similar to the performance of the new plan increases, and the performance data is suitable as a prediction of the performance. The processing device 1 calculates the second evaluation value of the performance data classified as each class according to the order of the ranking.
- the processing device 1 calculates the score of each set of performance data by using the first and second evaluation values. For example, the processing device 1 uses the sum of the first evaluation value and the second evaluation value, the product of the first evaluation value and the second evaluation value, or the average of the first evaluation value and the second evaluation value as the score. The processing device 1 may use a weighted sum of the first evaluation value and the second evaluation value as the score.
- the processing device 1 calculates a weight for each set of performance data.
- the weight for the performance data can be set based on the time difference between the start time of the new plan and the start time of the previous plan forming the basis of the performance data. It is considered that the performance data is more suitable as a prediction of the performance with respect to the new plan as the time difference decreases.
- the processing device 1 sets the weight to increase as the time difference decreases.
- the weight for the performance data may be set based on the difference between the total amount of the target production volume of the new plan and the total amount of the target production volume of the previous plan forming the basis of the performance data. It is considered that the suitability of the performance data increases as the production volume difference decreases.
- the processing device 1 sets the weight to increase as the production volume difference decreases.
- the total amount of the production volume can be calculated by summing the production volume at each time for the production period from the start time to the end time.
- the weight for the performance data may be set based on the difference between the length of the production period of the new plan and the length of the production period of the previous plan forming the basis of the performance data. It is considered that the suitability of the performance data increases as the production period difference decreases.
- the processing device 1 sets the weight to increase as the production period difference decreases.
- the processing device 1 extracts one set of performance data from the multiple sets of performance data based on the score. When a higher score indicates that the performance data is more suitable, the processing device 1 extracts the performance data having the highest score. The processing device 1 may extract two or more sets of performance data in order from the highest score. The processing device 1 stores the extracted performance data in the memory device 2 . The processing device 1 may output the extracted performance data to the output device 4 .
- the processing device 1 may generate first prediction data by correcting the production period and the production volume of the extracted performance data.
- the production period and the production volume can be corrected by the following method.
- the processing device 1 calculates the ratio of the production period of the performance data to the production period of the previous plan forming the basis of the performance data.
- the processing device 1 calculates the production period of the first prediction data by multiplying the ratio by the production period of the new plan.
- the processing device 1 sets the start time of the first prediction data to be the start time of the new plan.
- the processing device 1 sets the end time of the first prediction data by adding the calculated production period to the start time of the new plan.
- the processing device 1 calculates the ratio of the total amount of the actual production volume of the performance data to the total amount of the target production volume of the previous plan forming the basis of the performance data.
- the processing device 1 calculates the production volume of the first prediction data by multiplying the ratio by the target production volume at each time of the new plan.
- the processing device 1 stores the first prediction data in the memory device 2 .
- the processing device 1 may output the first prediction data to the output device 4 .
- FIGS. 5A and 5B , FIG. 6 , and FIGS. 7A and 7B are schematic views illustrating processing results according to the processing system according to the embodiment.
- the processing device 1 may output the processing results as illustrated in FIGS. 5A and 5B , FIG. 6 , and FIGS. 7A and 7B .
- the output method is arbitrary and may be a display by a monitor, a projection by a projector, printing by a printer, etc.
- a processing result 100 of FIG. 5A illustrates a result obtained by the first processing.
- one or more sets of previous plan data 111 of previous plans are displayed in a previous plan 110 column.
- One or more sets of performance data 121 of the performance with respect to the previous plans are displayed in a performance 120 column.
- the previous plan data 111 and the performance data 121 are displayed in a chart that the user can easily understand.
- the number of sets of the previous plan data 111 displayed in the previous plan 110 column and the number of sets of the performance data 121 displayed in the performance 120 column can be appropriately set by the user.
- Ranks 131 of each set of performance data when the performance data are ranked based on the first evaluation values are displayed in the rank 130 column.
- First evaluation values 141 in which each set of performance data is multiplied by the weight are displayed in an evaluation 140 column.
- a processing result 200 of FIG. 5B illustrates a result obtained by the second processing.
- one or more previous plan images 211 of the previous plans are displayed in a previous plan 210 column.
- One or more sets of performance data 221 of the performance with respect to the previous plans are displayed in a performance 220 column.
- the number of the previous plan images 211 displayed in the previous plan 210 column and the number of sets of the performance data 221 displayed in the performance 220 column can be appropriately set by the user.
- a performance image that shows the performance with respect to the previous plan may be displayed.
- Ranks 231 of each set of performance data when the performance data are ranked based on the second evaluation values are displayed in a rank 230 column.
- Second evaluation values 241 in which the weights are multiplied by each set of performance data are displayed in an evaluation 240 column.
- the processing device 1 calculates the score by summing the weighted first and second evaluation values.
- FIG. 6 illustrates a determination result obtained from the processing results of the first and second processing.
- the determination result 300 one or more sets of previous plan data 311 of the previous plans are displayed in a previous plan 310 column.
- a previous plan image may be displayed instead of the previous plan data.
- One or more sets of performance data 321 of the performance with respect to the previous plans are displayed in a performance 320 column.
- a performance image may be displayed instead of the performance data.
- the number of sets of the previous plan data 311 displayed in the previous plan 310 column and the number of sets of the performance data 321 displayed in the performance 320 column can be appropriately set by the user.
- Scores 341 that are the weighted sums of the first and second evaluation values for each set of performance data are displayed in an evaluation 340 column.
- Ranks 331 of each set of performance data when ranked based on the scores 341 are displayed in a rank 330 column.
- FIG. 7A illustrates an extraction result 400 .
- the extraction result 400 includes extracted performance data 420 and previous plan data 410 that forms the basis of the performance data 420 .
- the time of the production, the production volume, etc., are specifically displayed in the previous plan data 410 and the performance data 420 .
- the processing device 1 generates the first prediction data based on the extraction result.
- FIG. 7B illustrates a prediction result 500 .
- the prediction result 500 includes new plan data 510 and first prediction data 520 .
- the time of the production, the production volume, etc., are specifically displayed in the new plan data 510 and the first prediction data 520 .
- the new plan data 510 is prepared by the user and is data that is used in the first and second processing.
- the production period and the production volume of the first prediction data 520 are corrected using the previous plan data 410 , the performance data 420 , and the new plan data 510 .
- the user can easily ascertain how the processing device evaluates the performance data.
- the user can easily ascertain the performance data that is extracted and the scores of the performance data.
- the user can easily ascertain the data that is used in the first prediction data.
- the prediction result of FIG. 7B the user can easily ascertain the prediction result of the processing device 1 .
- the user can easily determine from a comparison of FIGS. 7A and 7B whether the data that is used in the first prediction data is appropriate.
- FIG. 8 is a flowchart illustrating a processing method according to the processing system according to the embodiment.
- the processing device 1 acquires the new plan data and the new plan image (step S 1 ).
- the processing device 1 executes the first processing (step S 2 ).
- the first evaluation value is obtained thereby.
- the processing device 1 weights the first evaluation value (step S 3 ).
- the processing device 1 executes the second processing (step S 4 ).
- the second evaluation value is obtained thereby.
- the processing device 1 weights the second evaluation value (step S 5 ).
- the processing device 1 calculates the score by using the weighted first and second evaluation values (step S 6 ).
- the processing device 1 extracts the performance data based on the score (step S 7 ).
- the processing device 1 generates the first prediction data (step S 8 ).
- the processing device 1 outputs the data obtained by the processing (step S 9 ).
- steps S 4 and S 5 may be executed before steps S 2 and S 3 .
- steps S 4 and S 5 may be executed in parallel with steps S 2 and S 3 .
- the production volume can be increased while using less resources; and more product types can be produced. As a result, for example, sales opportunities can be increased, and profits can be increased.
- the production volume may change over time. For example, the production volume changes according to demand fluctuation of products, product type increases due to the production of new products, etc.
- the necessary resources also change.
- the procurement and the distribution of resources are determined by planning the future relationship between time and the target production volume. At this time, it is desirable to determine the procurement and the distribution of resources while referring to the actual manufacturing capacity.
- the manufacturing capacity is an index that indicates the production volume for the allocated resources. By referring to the manufacturing capacity, excessive or insufficient resources can be reduced when producing according to the plan; and the resources can be more effectively utilized.
- the manufacturing capacity is calculated based on a rule or an optimization method that is set beforehand.
- uncertain elements exist in the actual manufacturing capacity.
- Unexpected absenteeism of workers, compatibility between the products and machines, aging of machines, fluctuation of worker skills, etc. are examples of uncertain elements.
- adaptation for these elements is performed in various ways based on skill and know-how such as modifying the personnel distribution, production adjustment of other products, etc.
- the knowledge, skill, experience, etc., of humans is necessary to produce products according to a new plan that is predetermined by a conventional method.
- time is necessary in order to investigate and execute the adaptation. Therefore, it is desirable to provide technology that can reduce the dependent on the knowledge, the skill, the experience, etc., of humans and can reduce the load of making the adaptation.
- data for estimating the actual manufacturing capacity with higher accuracy is provided to the user.
- the actual manufacturing capacity can be estimated with higher accuracy. Even when uncertain elements occur, a new plan that has a high likelihood of being executable can be generated.
- the processing system extracts not less than one set of performance data that can be utilized to predict the new plan from previous multiple sets of performance data.
- the extracted performance data is of the performance based on a previous plan that is similar to the new plan.
- the performance that is based on the previous plan also has a high likelihood of being similar to the performance based on the new plan.
- the performance data is a result in which uncertain elements have occurred in the previous plan and is of the actual manufacturing capacity. By referring to the extracted performance data, the actual manufacturing capacity can be estimated with higher accuracy.
- the performance data is extracted using the first and second evaluation values.
- the first evaluation value is calculated using the distance between the new plan data and the previous plan data.
- the second evaluation value is calculated using the classification result of the new plan image.
- FIG. 9 is a schematic view for describing advantages of embodiments.
- FIG. 9 illustrates multiple charts Ch 1 to Ch 3 showing relationships between time and the production volume.
- the horizontal axis is time; and the vertical axis is the production volume.
- the solid line illustrates the new plan.
- the dashed line illustrates the previous plan.
- the distance between the new plan and the previous plan illustrated in the chart Ch 3 is greater than the distance between the new plan and the previous plan illustrated in the chart Ch 2 . Therefore, it is determined that the new plan of the chart Ch 3 is not similar to the previous plan of the chart Ch 3 .
- the new plan is classified mainly according to the form of the change of the production volume with respect to time even when the difference between the target production volume of the new plan and the production volume of the performance data of the classified class is large. For example, even when the difference between the target production volume of the new plan and the target production volume of the previous plan is large as illustrated in the chart Ch 3 , the new plan is classified as the class of the performance data based on the previous plan. By using both the distance and the classification result of the image, performance data that is more useful for predicting the performance can be extracted.
- DTW it is favorable to use DTW when calculating the distance between the new plan data and the previous plan data. According to DTW, the change of the target production volume with respect to time can be compared between the new plan data and the previous plan data regardless of the length of the production period of each previous plan. Therefore, more useful performance data can be extracted.
- the first model it is favorable for the first model to include a CNN.
- a CNN the new plan that is used as the image can be more appropriately classified.
- the new plan can be more appropriately classified even when there are fine differences between the new plan and the previous plan used for the training and when there are abnormal values in the previous plan.
- the new plan can be classified as the class of the performance data based on a previous plan that has a general appearance similar to that of the new plan.
- the time of one or both of the previous plan data and the performance data may be corrected.
- the processing device 1 shifts the time of the performance data into the past by the amount of the time necessary to produce the product. For example, the time that is necessary for the production is stored in the previous plan as the lead time. Or, the processing device 1 may shift the time of the previous plan data into the future by the amount of the lead time.
- the processing device 1 may further execute third processing.
- the processing device 1 generates second prediction data based on production data.
- the production data is time-series data of the relationship between time and a production parameter.
- the production parameter includes at least one selected from the group consisting of an operation rate of equipment, an occurrence rate of discrepancies, a maintenance rate, and an occurrence rate of defective components.
- the processing device 1 When generating the plan, the change of the production parameter with respect to time also is predicted.
- the processing device 1 inputs the predicted production data to a second model.
- the second model outputs second prediction data of the relationship between time and the production volume.
- the second model includes a neural network. It is favorable for the second model to include a recurrent neural network (RNN). More favorably, the RNN includes a long short term memory (LSTM) structure.
- RNN recurrent neural network
- LSTM long short term memory
- the second model is pretrained.
- the performance data and the previous production data predicted when generating the previous plan are used to train the second model.
- the previous production data is time-series data of the relationship between time and a production parameter of the previous plan.
- the second model uses the previous production data as input data and is trained using the performance data as teaching data. Thereby, the second model is trained to be able to predict the performance data based on the production data.
- the processing device 1 stores the second prediction data in the memory device 2 .
- the processing device 1 may output the second prediction data to the output device 4 .
- FIG. 10 is a flowchart illustrating a processing method according to the processing system according to the modification of the embodiment.
- the processing device 1 acquires the new plan data, the new plan image, and the production data (step S 11 ). Similarly to the processing method PM 0 illustrated in FIG. 8 , the processing device 1 executes the first processing (step S 2 ) and weights the first evaluation value (step S 3 ). The processing device 1 executes the second processing (step S 4 ) and weights the second evaluation value (step S 5 ). The processing device 1 calculates the score (step S 6 ), extracts the performance data (step S 7 ), and generates the first prediction data (step S 8 ). The processing device 1 executes the third processing (step S 12 ). In the third processing, the processing device 1 acquires the second prediction data output from the second model. The processing device 1 outputs the data obtained by the processing (step S 9 ).
- the user can estimate the manufacturing capacity by referring to both the first and second prediction data. Even when a previous plan that is similar to the new plan does not exist, the user can utilize the second prediction data to predict the performance. According to the modification, the convenience of the user can be improved.
- FIG. 11 is a schematic view illustrating a hardware configuration.
- a computer 90 illustrated in FIG. 11 includes a CPU 91 , ROM 92 , RAM 93 , a memory device 94 , an input interface 95 , an output interface 96 , and a communication interface 97 .
- the ROM 92 stores programs that control the operations of the computer 90 .
- a program that is necessary for causing the computer 90 to realize the processing described above is stored in the ROM 92 .
- the RAM 93 functions as a memory region into which the programs stored in the ROM 92 are loaded.
- the CPU 91 includes a processing circuit.
- the CPU 91 uses the RAM 93 as work memory to execute the programs stored in at least one of the ROM 92 or the memory device 94 . While executing the program, the CPU 91 executes various processes by controlling configurations via a system bus 98 .
- the memory device 94 stores data necessary for executing the programs and data obtained by executing the programs.
- the input interface (I/F) 95 connects the computer 90 and an input device 95 a .
- the input I/F 95 is, for example, a serial bus interface such as USB, etc.
- the CPU 91 can read various data from the input device 95 a via the input I/F 95 .
- the output interface (I/F) 96 connects the computer 90 and an output device 96 a .
- the output I/F 96 is, for example, an image output interface such as Digital Visual Interface (DVI), High-Definition Multimedia Interface (HDMI (registered trademark)), etc.
- the CPU 91 can transmit the data to the output device 96 a via the output I/F 96 and can cause the output device 96 a to output the data.
- the output device 96 a outputs at least one of the processing results illustrated in FIGS. 5A and 5B , FIG. 6 , and FIGS. 7A and 7B .
- the communication interface (I/F) 97 connects the computer 90 and a server 97 a that is outside the computer 90 .
- the communication I/F 97 is, for example, a network card such as a LAN card, etc.
- the CPU 91 can read various data from the server 97 a via the communication I/F 97 .
- the memory device 94 includes not less than one selected from a hard disk drive (HDD) and a solid state drive (SSD).
- the input device 95 a includes not less than one selected from a mouse, a keyboard, a microphone (audio input), and a touchpad.
- the output device 96 a includes not less than one selected from a monitor, a projector, and a printer. A device such as a touch panel that functions as both the input device 95 a and the output device 96 a may be used.
- the computer 90 functions as the processing device 1 .
- the memory device 94 and the server 97 a function as the memory device 2 .
- the input device 95 a is used as the input device 3 of the processing system 10 .
- the output device 96 a is used as the output device 4 of the processing system 10 .
- data that is useful for generating the production plan can be provided to the user. Similar effects can be obtained by using a program to cause a computer to operate as the processing device.
- the processing of the various data described above may be recorded, as a program that can be executed by a computer, in a magnetic disk (a flexible disk, a hard disk, etc.), an optical disk (CD-ROM, CD-R, CD-RW, DVD-ROM, DVD ⁇ R, DVD ⁇ RW, etc.), semiconductor memory, or a recording medium (a non-transitory computer-readable storage medium) that can be read by another nontemporary computer.
- a magnetic disk a flexible disk, a hard disk, etc.
- an optical disk CD-ROM, CD-R, CD-RW, DVD-ROM, DVD ⁇ R, DVD ⁇ RW, etc.
- semiconductor memory or a recording medium (a non-transitory computer-readable storage medium) that can be read by another nontemporary computer.
- information that is recorded in the recording medium can be read by a computer (or an embedded system).
- the recording format (the storage format) of the recording medium is arbitrary.
- the computer reads the program from the recording medium and causes the CPU to execute the instructions recited in the program based on the program.
- the acquisition (or the reading) of the program may be performed via a network.
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Tourism & Hospitality (AREA)
- Educational Administration (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biodiversity & Conservation Biology (AREA)
- Manufacturing & Machinery (AREA)
- Primary Health Care (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
- This application is based upon and claims the benefit of priority from Japanese Patent Application No. 2020-206739, filed on Dec. 14, 2020; the entire contents of which are incorporated herein by reference.
- Embodiments described herein relate generally to a processing device, a processing method, and a storage medium.
- A production plan is pre-generated when producing a product. It is desirable to develop technology that can provide a user with data that is useful for generating the production plan.
-
FIG. 1 is a schematic view illustrating a processing system according to an embodiment; -
FIGS. 2A to 2H are schematic views for describing processing according to the processing system according to the embodiment; -
FIGS. 3A and 3B are schematic views for describing processing according to the processing system according to the embodiment; -
FIGS. 4A to 4D are schematic views for describing processing according to the processing system according to the embodiment; -
FIGS. 5A and 5B are schematic views illustrating processing results according to the processing system according to the embodiment; -
FIG. 6 is a schematic view illustrating processing results according to the processing system according to the embodiment; -
FIGS. 7A and 7B are schematic views illustrating processing results according to the processing system according to the embodiment; -
FIG. 8 is a flowchart illustrating a processing method according to the processing system according to the embodiment; -
FIG. 9 is a schematic view for describing advantages of embodiments; -
FIG. 10 is a flowchart illustrating a processing method according to the processing system according to the modification of the embodiment; and -
FIG. 11 is a schematic view illustrating a hardware configuration. - According to one embodiment, a processing device refers to a plurality of data sets. Each of the data sets includes previous plan data and performance data. The previous plan data are of a time series of a relationship between time and a target production volume of a previous plan. The performance data are of performance with respect to the previous plan. The device calculates a plurality of first evaluation values for a plurality of sets of the performance data by using distances between new plan data and each of a plurality of sets of the previous plan data. The new plan data are of a time series of a relationship between time and a target production volume of a new plan. The first evaluation values are of an evaluation as a prediction of performance with respect to the new plan data. The device inputs a new plan image to a first model and calculates a plurality of second evaluation values for the sets of performance data by using a classification result of the new plan image output from the first model. The new plan image is of the relationship between time and the target production volume of the new plan. The second evaluation values are of an evaluation as a prediction of the performance with respect to the new plan data. The device extracts at least one of the sets of performance data by using the first evaluation values and the second evaluation values.
- Various embodiments are described below with reference to the accompanying drawings. In the specification and drawings, components similar to those described previously or illustrated in an antecedent drawing are marked with like reference numerals, and a detailed description is omitted as appropriate.
-
FIG. 1 is a schematic view illustrating a processing system according to an embodiment. -
FIGS. 2A to 4D are schematic views for describing processing according to the processing system according to the embodiment. - The
processing system 10 according to the embodiment provides a user with data that can be utilized in a prediction of performance relating to a new production plan. Theprocessing system 10 includes aprocessing device 1, amemory device 2, aninput device 3, and anoutput device 4. - The
memory device 2 stores a data set that includes previous plan data and performance data for the previous plan data. The previous plan data is time-series data of the relationship between time and the target production volume at each time of a previous plan for some product. The performance data is time-series data of the relationship between time and the actual volume produced (the actual production volume) at each time for the product. For example, the previous plan data is of the relationship between dates and the target production volume of the product at each date. The performance data is of the relationship between dates and the actual production volume of the product at each date. - The
memory device 2 stores multiple data sets. The products that are the objects of the plans and performance may be the same or different from each other between the data sets. -
FIGS. 2A to 2D are examples of previous plan data. The previous plan data A1 to D1 ofFIGS. 2A to 2D respectively show the plans of the production of products A to D.FIGS. 2E to 2H are examples of performance data. The performance data A2 to D2 ofFIGS. 2E to 2H respectively show the performance with respect to the previous plan data A1 to D1 ofFIGS. 2A to 2D . InFIGS. 2A to 2H , the horizontal axis is time; and the vertical axis is the production volume. - A data set that includes the previous plan data A1 of
FIG. 2A and the performance data A2 ofFIG. 2E is stored in thememory device 2. Similarly, a data set of the previous plan data B1 ofFIG. 2B and the performance data B2 ofFIG. 2F , a data set of the previous plan data C1 ofFIG. 2C and the performance data C2 ofFIG. 2G , and a data set of the previous plan data D1 ofFIG. 2D and the performance data D2 ofFIG. 2H are stored in thememory device 2. - The
processing device 1 acquires new plan data. The new plan data is time-series data of the relationship between time and the target production volume at each time of a new plan for any product. For example, the new plan data is of the relationship between future dates and the target production volume of the product at each date. The product that is the object of the new plan may be the same as or different from the products that are the objects of the previous plans and performance. - The user uses the
input device 3 to input data to theprocessing device 1. The data that is transmitted from theprocessing device 1 is output by theoutput device 4 toward the user. For example, the user uses theinput device 3 to input the new plan data to theprocessing device 1. The new plan data is stored in thememory device 2; and theprocessing device 1 may acquire the new plan data from thememory device 2. -
FIG. 3A is an example of new plan data. The new plan data N1 ofFIG. 3A shows the plan of the production of the product A. InFIG. 3A , the horizontal axis is time; and the vertical axis is the production volume. - The
processing device 1 extracts performance data that can be utilized as a prediction of the performance with respect to the new plan from the multiple data sets. Theprocessing device 1 extracts by executing the following first processing and second processing. - In the first processing, the
processing device 1 calculates the distances between the new plan data and the multiple sets of previous plan data. Theprocessing device 1 calculates the multiple first evaluation values of an evaluation as a prediction of the performance with respect to the new plan data for the multiple sets of performance data by using the calculated multiple distances. - For example, the
processing device 1 calculates the distances between the new plan data N1 ofFIG. 3A and the multiple sets of previous plan data A1 to D1 ofFIGS. 2A to 2D . The distance decreases as the similarity of the form of the change of the target production volume with respect to time increases. The performance data for the previous plan data for which the distance is obtained becomes more suitable to predict the performance with respect to the new plan data as the distance decreases. - The
processing device 1 generates a ranking by arranging the multiple sets of performance data in order from the shortest distance for the multiple distances of the multiple sets of previous plan data. The ranking of the performance data corresponding to the previous plan data is higher as the distance for the previous plan data decreases. Theprocessing device 1 calculates the first evaluation values of the performance data according to the order of the ranking. - A Euclidean distance, a Manhattan distance, a Chebyshev distance, a distance calculated by dynamic time warping (DTW), etc., can be used as the distance.
- In the second processing, the
processing device 1 uses the new plan image and a first model. The new plan image is image data of the relationship between time and the target production volume of the new plan.FIG. 3B is an example of a new plan image. The new plan image n1 ofFIG. 3B is image data corresponding to the new plan data N1 ofFIG. 3A . - In the example, the image of a chart in which the horizontal axis is time and the vertical axis is the production volume is used as the new plan image n1. The content shown in the image is arbitrary as long as the relationship between time and the target production volume is shown. For example, the image may be a scatter plot showing the relationship between time and the production volume.
- The user uses the
input device 3 to input the new plan image to theprocessing device 1. Or, theprocessing device 1 may generate the new plan image based on the new plan data. - The first model outputs a classification result according to the input of the plan image. The first model includes, for example, a neural network, a model trained by a random forest, a model trained by a decision tree, or a model trained by a support vector machine. It is favorable for the first model to include a convolutional neural network (CNN). Each of the multiple sets of performance data are classified as some class. The classification result includes the probability of the plan image being classified as each of the multiple classes.
- The first model is pretrained using a previous plan image. The previous plan image is image data of the relationship between time and the target production volume of the previous plan for some product.
-
FIGS. 4A to 4D are examples of previous plan images. The previous plan images al to dl ofFIGS. 4A to 4D correspond respectively to the previous plan data A1 to D1 ofFIGS. 2A to 2D and show plans of the production of the products A to D. In the example, the image of a chart in which the horizontal axis is time and the vertical axis is the production volume is used as the previous plan image. The performance data A2 to D2 ofFIGS. 2E to 2H respectively show the performance with respect to the previous plans of the previous plan images al to dl. - The first model is trained by using the previous plan image as input data and by using a value of the class of the performance data as a label. For example, the first model is trained by using the previous plan image al of
FIG. 4A as input data and by using a value of a class that includes the performance data A2 ofFIG. 2E as a label. Thereby, the first model is trained so that the image is classified as some class when the plan image is input. - The class is set according to the form of the change of the target production volume with respect to time. For example, the same class is set for previous plans that have similar forms of the change of the target production volume with respect to time regardless of the target production volume and the production period.
- The
processing device 1 inputs the new plan image to the first model and acquires a classification result output from the first model. Theprocessing device 1 uses the classification result to calculate a second evaluation value for each set of performance data. The second evaluation value is of an evaluation as a prediction of the performance with respect to the new plan data of the new plan image for the performance data. - For example, the
processing device 1 generates a ranking by arranging the classes in order from highest probability. As the probability increases, the likelihood of the performance data classified as each class being similar to the performance of the new plan increases, and the performance data is suitable as a prediction of the performance. Theprocessing device 1 calculates the second evaluation value of the performance data classified as each class according to the order of the ranking. - The
processing device 1 calculates the score of each set of performance data by using the first and second evaluation values. For example, theprocessing device 1 uses the sum of the first evaluation value and the second evaluation value, the product of the first evaluation value and the second evaluation value, or the average of the first evaluation value and the second evaluation value as the score. Theprocessing device 1 may use a weighted sum of the first evaluation value and the second evaluation value as the score. - For example, the
processing device 1 calculates a weight for each set of performance data. The weight for the performance data can be set based on the time difference between the start time of the new plan and the start time of the previous plan forming the basis of the performance data. It is considered that the performance data is more suitable as a prediction of the performance with respect to the new plan as the time difference decreases. Theprocessing device 1 sets the weight to increase as the time difference decreases. - The weight for the performance data may be set based on the difference between the total amount of the target production volume of the new plan and the total amount of the target production volume of the previous plan forming the basis of the performance data. It is considered that the suitability of the performance data increases as the production volume difference decreases. The
processing device 1 sets the weight to increase as the production volume difference decreases. The total amount of the production volume can be calculated by summing the production volume at each time for the production period from the start time to the end time. - Or, the weight for the performance data may be set based on the difference between the length of the production period of the new plan and the length of the production period of the previous plan forming the basis of the performance data. It is considered that the suitability of the performance data increases as the production period difference decreases. The
processing device 1 sets the weight to increase as the production period difference decreases. - The
processing device 1 extracts one set of performance data from the multiple sets of performance data based on the score. When a higher score indicates that the performance data is more suitable, theprocessing device 1 extracts the performance data having the highest score. Theprocessing device 1 may extract two or more sets of performance data in order from the highest score. Theprocessing device 1 stores the extracted performance data in thememory device 2. Theprocessing device 1 may output the extracted performance data to theoutput device 4. - The
processing device 1 may generate first prediction data by correcting the production period and the production volume of the extracted performance data. The production period and the production volume can be corrected by the following method. - The
processing device 1 calculates the ratio of the production period of the performance data to the production period of the previous plan forming the basis of the performance data. Theprocessing device 1 calculates the production period of the first prediction data by multiplying the ratio by the production period of the new plan. Theprocessing device 1 sets the start time of the first prediction data to be the start time of the new plan. Theprocessing device 1 sets the end time of the first prediction data by adding the calculated production period to the start time of the new plan. - The
processing device 1 calculates the ratio of the total amount of the actual production volume of the performance data to the total amount of the target production volume of the previous plan forming the basis of the performance data. Theprocessing device 1 calculates the production volume of the first prediction data by multiplying the ratio by the target production volume at each time of the new plan. - The
processing device 1 stores the first prediction data in thememory device 2. Theprocessing device 1 may output the first prediction data to theoutput device 4. -
FIGS. 5A and 5B ,FIG. 6 , andFIGS. 7A and 7B are schematic views illustrating processing results according to the processing system according to the embodiment. - The
processing device 1 may output the processing results as illustrated inFIGS. 5A and 5B ,FIG. 6 , andFIGS. 7A and 7B . The output method is arbitrary and may be a display by a monitor, a projection by a projector, printing by a printer, etc. - A
processing result 100 ofFIG. 5A illustrates a result obtained by the first processing. In theprocessing result 100, one or more sets ofprevious plan data 111 of previous plans are displayed in aprevious plan 110 column. One or more sets ofperformance data 121 of the performance with respect to the previous plans are displayed in aperformance 120 column. Theprevious plan data 111 and theperformance data 121 are displayed in a chart that the user can easily understand. The number of sets of theprevious plan data 111 displayed in theprevious plan 110 column and the number of sets of theperformance data 121 displayed in theperformance 120 column can be appropriately set by the user.Ranks 131 of each set of performance data when the performance data are ranked based on the first evaluation values are displayed in therank 130 column. First evaluation values 141 in which each set of performance data is multiplied by the weight are displayed in anevaluation 140 column. - A
processing result 200 ofFIG. 5B illustrates a result obtained by the second processing. In theprocessing result 200, one or moreprevious plan images 211 of the previous plans are displayed in aprevious plan 210 column. One or more sets ofperformance data 221 of the performance with respect to the previous plans are displayed in aperformance 220 column. The number of theprevious plan images 211 displayed in theprevious plan 210 column and the number of sets of theperformance data 221 displayed in theperformance 220 column can be appropriately set by the user. Instead of the performance data, a performance image that shows the performance with respect to the previous plan may be displayed.Ranks 231 of each set of performance data when the performance data are ranked based on the second evaluation values are displayed in arank 230 column. Second evaluation values 241 in which the weights are multiplied by each set of performance data are displayed in anevaluation 240 column. - The
processing device 1 calculates the score by summing the weighted first and second evaluation values.FIG. 6 illustrates a determination result obtained from the processing results of the first and second processing. In thedetermination result 300, one or more sets ofprevious plan data 311 of the previous plans are displayed in aprevious plan 310 column. A previous plan image may be displayed instead of the previous plan data. One or more sets ofperformance data 321 of the performance with respect to the previous plans are displayed in aperformance 320 column. A performance image may be displayed instead of the performance data. The number of sets of theprevious plan data 311 displayed in theprevious plan 310 column and the number of sets of theperformance data 321 displayed in theperformance 320 column can be appropriately set by the user.Scores 341 that are the weighted sums of the first and second evaluation values for each set of performance data are displayed in anevaluation 340 column. Ranks 331 of each set of performance data when ranked based on thescores 341 are displayed in a rank 330 column. - For example, the
processing device 1 extracts the performance data having the highest score and the previous plan data forming the basis of the performance data.FIG. 7A illustrates anextraction result 400. Theextraction result 400 includes extractedperformance data 420 andprevious plan data 410 that forms the basis of theperformance data 420. The time of the production, the production volume, etc., are specifically displayed in theprevious plan data 410 and theperformance data 420. - The
processing device 1 generates the first prediction data based on the extraction result.FIG. 7B illustrates aprediction result 500. Theprediction result 500 includesnew plan data 510 andfirst prediction data 520. The time of the production, the production volume, etc., are specifically displayed in thenew plan data 510 and thefirst prediction data 520. Thenew plan data 510 is prepared by the user and is data that is used in the first and second processing. The production period and the production volume of thefirst prediction data 520 are corrected using theprevious plan data 410, theperformance data 420, and thenew plan data 510. - By confirming a first ranking in which not less than one of the multiple sets of performance data illustrated in
FIG. 5A are arranged and a second ranking in which not less than one of the multiple sets of performance data illustrated inFIG. 5B are arranged, the user can easily ascertain how the processing device evaluates the performance data. By confirming the determination result ofFIG. 6 , the user can easily ascertain the performance data that is extracted and the scores of the performance data. By confirming the extraction result ofFIG. 7A , the user can easily ascertain the data that is used in the first prediction data. By confirming the prediction result ofFIG. 7B , the user can easily ascertain the prediction result of theprocessing device 1. For example, the user can easily determine from a comparison ofFIGS. 7A and 7B whether the data that is used in the first prediction data is appropriate. -
FIG. 8 is a flowchart illustrating a processing method according to the processing system according to the embodiment. - In the processing method PM0, the
processing device 1 acquires the new plan data and the new plan image (step S1). Theprocessing device 1 executes the first processing (step S2). The first evaluation value is obtained thereby. Theprocessing device 1 weights the first evaluation value (step S3). Theprocessing device 1 executes the second processing (step S4). The second evaluation value is obtained thereby. Theprocessing device 1 weights the second evaluation value (step S5). Theprocessing device 1 calculates the score by using the weighted first and second evaluation values (step S6). Theprocessing device 1 extracts the performance data based on the score (step S7). Theprocessing device 1 generates the first prediction data (step S8). Theprocessing device 1 outputs the data obtained by the processing (step S9). - In the processing method PM0, the order of the processing is modifiable as appropriate. For example, steps S4 and S5 may be executed before steps S2 and S3. Steps S4 and S5 may be executed in parallel with steps S2 and S3.
- Advantages of embodiments will now be described.
- In the production of the product, it is desirable to more effectively utilize limited resources. By effectively utilizing resources, the production volume can be increased while using less resources; and more product types can be produced. As a result, for example, sales opportunities can be increased, and profits can be increased. On the other hand, the production volume may change over time. For example, the production volume changes according to demand fluctuation of products, product type increases due to the production of new products, etc. When the production volume changes, the necessary resources also change. To adapt to the change of the production volume, the procurement and the distribution of resources are determined by planning the future relationship between time and the target production volume. At this time, it is desirable to determine the procurement and the distribution of resources while referring to the actual manufacturing capacity. The manufacturing capacity is an index that indicates the production volume for the allocated resources. By referring to the manufacturing capacity, excessive or insufficient resources can be reduced when producing according to the plan; and the resources can be more effectively utilized.
- Conventionally, the manufacturing capacity is calculated based on a rule or an optimization method that is set beforehand. However, many uncertain elements exist in the actual manufacturing capacity. Unexpected absenteeism of workers, compatibility between the products and machines, aging of machines, fluctuation of worker skills, etc., are examples of uncertain elements. As a result, it is difficult to produce according to a new plan that is predetermined. Conventionally, adaptation for these elements is performed in various ways based on skill and know-how such as modifying the personnel distribution, production adjustment of other products, etc. In other words, the knowledge, skill, experience, etc., of humans is necessary to produce products according to a new plan that is predetermined by a conventional method. Also, time is necessary in order to investigate and execute the adaptation. Therefore, it is desirable to provide technology that can reduce the dependent on the knowledge, the skill, the experience, etc., of humans and can reduce the load of making the adaptation.
- For this problem, according to the embodiment, data for estimating the actual manufacturing capacity with higher accuracy is provided to the user. By referring to the data, the actual manufacturing capacity can be estimated with higher accuracy. Even when uncertain elements occur, a new plan that has a high likelihood of being executable can be generated.
- Specifically, the processing system according to the embodiment extracts not less than one set of performance data that can be utilized to predict the new plan from previous multiple sets of performance data. The extracted performance data is of the performance based on a previous plan that is similar to the new plan. When the previous plan is similar to the new plan, the performance that is based on the previous plan also has a high likelihood of being similar to the performance based on the new plan. The performance data is a result in which uncertain elements have occurred in the previous plan and is of the actual manufacturing capacity. By referring to the extracted performance data, the actual manufacturing capacity can be estimated with higher accuracy.
- According to embodiments, the performance data is extracted using the first and second evaluation values. The first evaluation value is calculated using the distance between the new plan data and the previous plan data. The second evaluation value is calculated using the classification result of the new plan image. By using the distance and the classification result of the image, a previous plan that is similar to the new plan can be extracted. As a result, performance data that is useful for predicting the performance with respect to the new plan can be extracted.
-
FIG. 9 is a schematic view for describing advantages of embodiments. -
FIG. 9 illustrates multiple charts Ch1 to Ch3 showing relationships between time and the production volume. In the charts Ch1 to Ch3, the horizontal axis is time; and the vertical axis is the production volume. The solid line illustrates the new plan. The dashed line illustrates the previous plan. By using the distance, a previous plan can be extracted so that the target production volume difference at each time between the previous plan and the new plan is small as illustrated in the charts Ch1 and Ch2. On the other hand, when extracting by using the distance, it is difficult to extract a previous plan when the target production volume difference between the previous plan and the new plan is large even though the forms of the change of the target production volume with respect to time are similar. For example, the distance between the new plan and the previous plan illustrated in the chart Ch3 is greater than the distance between the new plan and the previous plan illustrated in the chart Ch2. Therefore, it is determined that the new plan of the chart Ch3 is not similar to the previous plan of the chart Ch3. - When the classification result of the image is used, the new plan is classified mainly according to the form of the change of the production volume with respect to time even when the difference between the target production volume of the new plan and the production volume of the performance data of the classified class is large. For example, even when the difference between the target production volume of the new plan and the target production volume of the previous plan is large as illustrated in the chart Ch3, the new plan is classified as the class of the performance data based on the previous plan. By using both the distance and the classification result of the image, performance data that is more useful for predicting the performance can be extracted.
- It is favorable to use DTW when calculating the distance between the new plan data and the previous plan data. According to DTW, the change of the target production volume with respect to time can be compared between the new plan data and the previous plan data regardless of the length of the production period of each previous plan. Therefore, more useful performance data can be extracted.
- It is favorable for the first model to include a CNN. According to a CNN, the new plan that is used as the image can be more appropriately classified. For example, the new plan can be more appropriately classified even when there are fine differences between the new plan and the previous plan used for the training and when there are abnormal values in the previous plan. In other words, the new plan can be classified as the class of the performance data based on a previous plan that has a general appearance similar to that of the new plan.
- The time of one or both of the previous plan data and the performance data may be corrected. The
processing device 1 shifts the time of the performance data into the past by the amount of the time necessary to produce the product. For example, the time that is necessary for the production is stored in the previous plan as the lead time. Or, theprocessing device 1 may shift the time of the previous plan data into the future by the amount of the lead time. - The
processing device 1 may further execute third processing. In the third processing, theprocessing device 1 generates second prediction data based on production data. The production data is time-series data of the relationship between time and a production parameter. The production parameter includes at least one selected from the group consisting of an operation rate of equipment, an occurrence rate of discrepancies, a maintenance rate, and an occurrence rate of defective components. - When generating the plan, the change of the production parameter with respect to time also is predicted. When generating the new plan, the
processing device 1 inputs the predicted production data to a second model. The second model outputs second prediction data of the relationship between time and the production volume. - The second model includes a neural network. It is favorable for the second model to include a recurrent neural network (RNN). More favorably, the RNN includes a long short term memory (LSTM) structure.
- The second model is pretrained. The performance data and the previous production data predicted when generating the previous plan are used to train the second model. The previous production data is time-series data of the relationship between time and a production parameter of the previous plan. The second model uses the previous production data as input data and is trained using the performance data as teaching data. Thereby, the second model is trained to be able to predict the performance data based on the production data.
- The
processing device 1 stores the second prediction data in thememory device 2. Theprocessing device 1 may output the second prediction data to theoutput device 4. -
FIG. 10 is a flowchart illustrating a processing method according to the processing system according to the modification of the embodiment. - In the processing method PM1, the
processing device 1 acquires the new plan data, the new plan image, and the production data (step S11). Similarly to the processing method PM0 illustrated inFIG. 8 , theprocessing device 1 executes the first processing (step S2) and weights the first evaluation value (step S3). Theprocessing device 1 executes the second processing (step S4) and weights the second evaluation value (step S5). Theprocessing device 1 calculates the score (step S6), extracts the performance data (step S7), and generates the first prediction data (step S8). Theprocessing device 1 executes the third processing (step S12). In the third processing, theprocessing device 1 acquires the second prediction data output from the second model. Theprocessing device 1 outputs the data obtained by the processing (step S9). - According to the modification, the user can estimate the manufacturing capacity by referring to both the first and second prediction data. Even when a previous plan that is similar to the new plan does not exist, the user can utilize the second prediction data to predict the performance. According to the modification, the convenience of the user can be improved.
-
FIG. 11 is a schematic view illustrating a hardware configuration. - The
processing system 10 according to the embodiment can be realized by the hardware configuration illustrated inFIG. 11 . A computer 90 illustrated inFIG. 11 includes aCPU 91,ROM 92,RAM 93, amemory device 94, aninput interface 95, anoutput interface 96, and acommunication interface 97. - The
ROM 92 stores programs that control the operations of the computer 90. A program that is necessary for causing the computer 90 to realize the processing described above is stored in theROM 92. TheRAM 93 functions as a memory region into which the programs stored in theROM 92 are loaded. - The
CPU 91 includes a processing circuit. TheCPU 91 uses theRAM 93 as work memory to execute the programs stored in at least one of theROM 92 or thememory device 94. While executing the program, theCPU 91 executes various processes by controlling configurations via asystem bus 98. - The
memory device 94 stores data necessary for executing the programs and data obtained by executing the programs. - The input interface (I/F) 95 connects the computer 90 and an
input device 95 a. The input I/F 95 is, for example, a serial bus interface such as USB, etc. TheCPU 91 can read various data from theinput device 95 a via the input I/F 95. - The output interface (I/F) 96 connects the computer 90 and an
output device 96 a. The output I/F 96 is, for example, an image output interface such as Digital Visual Interface (DVI), High-Definition Multimedia Interface (HDMI (registered trademark)), etc. TheCPU 91 can transmit the data to theoutput device 96 a via the output I/F 96 and can cause theoutput device 96 a to output the data. For example, theoutput device 96 a outputs at least one of the processing results illustrated inFIGS. 5A and 5B ,FIG. 6 , andFIGS. 7A and 7B . - The communication interface (I/F) 97 connects the computer 90 and a
server 97 a that is outside the computer 90. The communication I/F 97 is, for example, a network card such as a LAN card, etc. TheCPU 91 can read various data from theserver 97 a via the communication I/F 97. - The
memory device 94 includes not less than one selected from a hard disk drive (HDD) and a solid state drive (SSD). Theinput device 95 a includes not less than one selected from a mouse, a keyboard, a microphone (audio input), and a touchpad. Theoutput device 96 a includes not less than one selected from a monitor, a projector, and a printer. A device such as a touch panel that functions as both theinput device 95 a and theoutput device 96 a may be used. - The computer 90 functions as the
processing device 1. Thememory device 94 and theserver 97 a function as thememory device 2. Theinput device 95 a is used as theinput device 3 of theprocessing system 10. Theoutput device 96 a is used as theoutput device 4 of theprocessing system 10. - By using the processing device or the processing method described above, data that is useful for generating the production plan can be provided to the user. Similar effects can be obtained by using a program to cause a computer to operate as the processing device.
- The processing of the various data described above may be recorded, as a program that can be executed by a computer, in a magnetic disk (a flexible disk, a hard disk, etc.), an optical disk (CD-ROM, CD-R, CD-RW, DVD-ROM, DVD±R, DVD±RW, etc.), semiconductor memory, or a recording medium (a non-transitory computer-readable storage medium) that can be read by another nontemporary computer.
- For example, information that is recorded in the recording medium can be read by a computer (or an embedded system). The recording format (the storage format) of the recording medium is arbitrary. For example, the computer reads the program from the recording medium and causes the CPU to execute the instructions recited in the program based on the program. In the computer, the acquisition (or the reading) of the program may be performed via a network.
- While certain embodiments have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. Indeed, the novel embodiments described herein may be embodied in a variety of other forms; furthermore, various omissions, substitutions and changes in the form of the embodiments described herein may be made without departing from the spirit of the inventions. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the invention. The above embodiments can be practiced by combining data sets mutually.
Claims (13)
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2020-206739 | 2020-12-14 | ||
JP2020206739A JP2022093976A (en) | 2020-12-14 | 2020-12-14 | Processing device, processing methods, programs, and storage media |
Publications (2)
Publication Number | Publication Date |
---|---|
US20220188734A1 true US20220188734A1 (en) | 2022-06-16 |
US11900298B2 US11900298B2 (en) | 2024-02-13 |
Family
ID=81941553
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US17/408,696 Active 2041-09-25 US11900298B2 (en) | 2020-12-14 | 2021-08-23 | Processing device, processing method, and storage medium for predicting performance of a production plan |
Country Status (3)
Country | Link |
---|---|
US (1) | US11900298B2 (en) |
JP (1) | JP2022093976A (en) |
CN (1) | CN114691951A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5231567A (en) * | 1990-11-28 | 1993-07-27 | Hitachi, Ltd. | Manufacturing planning system |
US5479343A (en) * | 1990-11-28 | 1995-12-26 | Hitachi, Ltd. | Production planning system |
US5586021A (en) * | 1992-03-24 | 1996-12-17 | Texas Instruments Incorporated | Method and system for production planning |
US5657453A (en) * | 1992-02-14 | 1997-08-12 | Mitsubishi Denki Kabishiki Kaisha | Successively-deciding production planning system |
US20070244591A1 (en) * | 2006-04-04 | 2007-10-18 | Hitachi Global Storage Technologies Netherlands, B.V. | Production planning method and production planning system |
US20220152901A1 (en) * | 2020-11-17 | 2022-05-19 | Seiko Epson Corporation | Production management system |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000077289A (en) | 1998-08-27 | 2000-03-14 | Fujitsu Ltd | Production predicting control system |
KR101917006B1 (en) | 2016-11-30 | 2018-11-08 | 에스케이 주식회사 | Semiconductor Manufacturing Yield Prediction System and Method based on Machine Learning |
JP7030072B2 (en) | 2019-03-14 | 2022-03-04 | 株式会社日立製作所 | Time-series data monitoring system and time-series data monitoring method |
-
2020
- 2020-12-14 JP JP2020206739A patent/JP2022093976A/en active Pending
-
2021
- 2021-08-23 US US17/408,696 patent/US11900298B2/en active Active
- 2021-09-03 CN CN202111030492.5A patent/CN114691951A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5231567A (en) * | 1990-11-28 | 1993-07-27 | Hitachi, Ltd. | Manufacturing planning system |
US5479343A (en) * | 1990-11-28 | 1995-12-26 | Hitachi, Ltd. | Production planning system |
US5657453A (en) * | 1992-02-14 | 1997-08-12 | Mitsubishi Denki Kabishiki Kaisha | Successively-deciding production planning system |
US5586021A (en) * | 1992-03-24 | 1996-12-17 | Texas Instruments Incorporated | Method and system for production planning |
US20070244591A1 (en) * | 2006-04-04 | 2007-10-18 | Hitachi Global Storage Technologies Netherlands, B.V. | Production planning method and production planning system |
US7693593B2 (en) * | 2006-04-04 | 2010-04-06 | Hitachi Global Storage Technologies Netherlands B.V. | Production planning method and production planning system |
US20220152901A1 (en) * | 2020-11-17 | 2022-05-19 | Seiko Epson Corporation | Production management system |
Non-Patent Citations (2)
Title |
---|
Ha, Chunghun, Hyesung Seok, and Changsoo Ok. "Evaluation of forecasting methods in aggregate production planning: A Cumulative Absolute Forecast Error (CAFE)." Computers & Industrial Engineering 118 (2018): 329-339. (Year: 2018) * |
Vieira, Guilherme E., Jeffrey W. Herrmann, and Edward Lin. "Predicting the performance of rescheduling strategies for parallel machine systems." Journal of manufacturing Systems 19.4 (2000): 256-266. (Year: 2000) * |
Also Published As
Publication number | Publication date |
---|---|
JP2022093976A (en) | 2022-06-24 |
US11900298B2 (en) | 2024-02-13 |
CN114691951A (en) | 2022-07-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10970158B2 (en) | Root cause analysis | |
US10067936B2 (en) | Machine translation output reranking | |
JP6451895B2 (en) | Prediction model selection system, prediction model selection method, and prediction model selection program | |
US11093702B2 (en) | Checking and/or completion for data grids | |
EP3455741A1 (en) | Automated accuracy assessment in tasking system | |
JP5846442B2 (en) | Information processing apparatus, information processing method, and program | |
US20200320419A1 (en) | Method and device of classification models construction and data prediction | |
JP7155758B2 (en) | Information processing device, information processing method and program | |
US20180174090A1 (en) | Production Management Support Apparatus, Production Management Support Method, and Production Management Support Program | |
JP6668892B2 (en) | Item recommendation program, item recommendation method and item recommendation device | |
US10108879B2 (en) | Aggregate training data set generation for OCR processing | |
JP2007323315A (en) | Cooperative filtering method, cooperative filtering device, cooperative filtering program and recording medium with the same program recorded thereon | |
US20060156264A1 (en) | Method and apparatus for supporting verification of system, and computer product | |
US11900298B2 (en) | Processing device, processing method, and storage medium for predicting performance of a production plan | |
US20210248293A1 (en) | Optimization device and optimization method | |
JP7274434B2 (en) | Diversion design support system and diversion design support method | |
JP7005463B2 (en) | Learning equipment, learning methods and programs | |
JP7215318B2 (en) | Information processing program, information processing method, and information processing apparatus | |
US20220405534A1 (en) | Learning apparatus, information integration system, learning method, and recording medium | |
US20180253515A1 (en) | Characterizing model performance using hierarchical feature groups | |
JP2016099688A (en) | Risk evaluation method and risk evaluation device | |
JP7473245B2 (en) | Judgment method | |
JP7396762B1 (en) | Information processing system, information processing method, and program | |
JP2019133478A (en) | Computing system | |
JP7189922B2 (en) | Information processing device, information processing method and information processing program |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FEPP | Fee payment procedure |
Free format text: ENTITY STATUS SET TO UNDISCOUNTED (ORIGINAL EVENT CODE: BIG.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |
|
AS | Assignment |
Owner name: KABUSHIKI KAISHA TOSHIBA, JAPAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:NIINUMA, AYUMU;OSHIMA, HIROTOMO;REEL/FRAME:057849/0422 Effective date: 20211007 |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NON FINAL ACTION MAILED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NOTICE OF ALLOWANCE MAILED -- APPLICATION RECEIVED IN OFFICE OF PUBLICATIONS |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: PUBLICATIONS -- ISSUE FEE PAYMENT RECEIVED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: PUBLICATIONS -- ISSUE FEE PAYMENT VERIFIED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: AWAITING TC RESP, ISSUE FEE PAYMENT VERIFIED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: PUBLICATIONS -- ISSUE FEE PAYMENT VERIFIED |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |